An Investigation of ETF Flows : Asset Allocation Perspectives

Abstract

This study investigates impacts of asset allocations on ETF flows across asset classes in a panel form, utilizing fixed-effects panel regression. The period of study covers October 2008 to July 2019. The investment universe contains 9 U.S.-listed ETFs, classifying into 5 asset classes. Monthly asset allocations are obtained from 5 optimization strategies based on mean-variance and risk-based optimizations. The findings indicate that asset allocations of risk-based optimization strategies significantly explain fund flows across asset classes. It implies that ETF market participants attempt to stabilize portfolio volatilities rather than maximize portfolio expected returns or risk-adjusted returns. This study further analyses relations between ETF flows and returns during flight-to-quality episodes and find that flows lead returns.